SuperMicro A2SDi-4C-HLN4F

classic Classic list List threaded Threaded
19 messages Options
Reply | Threaded
Open this post in threaded view
|

SuperMicro A2SDi-4C-HLN4F

Rupert Gallagher
Reply | Threaded
Open this post in threaded view
|

Re: SuperMicro A2SDi-4C-HLN4F

Rupert Gallagher
Using the serial over lan console, the keyboard keeps working, and it is possible to enter the shell. Using the standard console, the keyboard does not work. There is plenty of drivers that need to be written. No, I do not write/port drivers.

By comparison, both freebsd and centos linux just work out of the box. Both dmesgs are attached as reference.

The board is really nice.

On Thu, Aug 16, 2018 at 18:04, Rupert Gallagher <[hidden email]> wrote:

> https://www.supermicro.com/products/motherboard/atom/A2SDi-4C-HLN4F.cfm
Reply | Threaded
Open this post in threaded view
|

Re: SuperMicro A2SDi-4C-HLN4F

lists-2
Fri, 17 Aug 2018 06:23:23 +0000 Rupert Gallagher <[hidden email]>
> Using the serial over lan console, the keyboard keeps working, and it is possible to enter the shell. Using the standard console, the keyboard does not work. There is plenty of drivers that need to be written. No, I do not write/port drivers.
>
> By comparison, both freebsd and centos linux just work out of the box. Both dmesgs are attached as reference.
>
> The board is really nice.
>
> On Thu, Aug 16, 2018 at 18:04, Rupert Gallagher <[hidden email]> wrote:
>
> > https://www.supermicro.com/products/motherboard/atom/A2SDi-4C-HLN4F.cfm 

Hi misc@,

This one is much quieter, the vendors have reacted to the previous flaw.
As nice as it can get.. before trusting non-technical reviews, be aware:

https://en.wikipedia.org/wiki/Goldmont#Erratum
https://www.servethehome.com/another-atom-bomb-intel-e3800-bay-trail-atom-vli89-bug/
https://www.google.com/search?q=intel+c3000+bug

Recurring issue - similar problems were discussed noisily industry wide:

https://en.wikipedia.org/wiki/Silvermont#Erratum
https://www.servethehome.com/intel-atom-c2000-series-bug-quiet/
https://www.servethehome.com/intel-atom-c2000-c0-stepping-fixing-the-avr54-bug/
https://www.google.com/search?q=intel+c2000+bug

As with the previous generation processors and main boards, Intel FAILS.
Wait for updated CPU revision boards, consult techs and prepare for RMA.

Kind regards,
Anton Lazarov

Reply | Threaded
Open this post in threaded view
|

Re: SuperMicro A2SDi-4C-HLN4F

Rupert Gallagher
FUD

Sent from ProtonMail Mobile

On Fri, Aug 17, 2018 at 12:24, <[hidden email]> wrote:

> Fri, 17 Aug 2018 06:23:23 +0000 Rupert Gallagher <[hidden email]>
>> Using the serial over lan console, the keyboard keeps working, and it is possible to enter the shell. Using the standard console, the keyboard does not work. There is plenty of drivers that need to be written. No, I do not write/port drivers.
>>
>> By comparison, both freebsd and centos linux just work out of the box. Both dmesgs are attached as reference.
>>
>> The board is really nice.
>>
>> On Thu, Aug 16, 2018 at 18:04, Rupert Gallagher <[hidden email]> wrote:
>>
>> > https://www.supermicro.com/products/motherboard/atom/A2SDi-4C-HLN4F.cfm
>
> Hi misc@,
>
> This one is much quieter, the vendors have reacted to the previous flaw.
> As nice as it can get.. before trusting non-technical reviews, be aware:
>
> https://en.wikipedia.org/wiki/Goldmont#Erratum
> https://www.servethehome.com/another-atom-bomb-intel-e3800-bay-trail-atom-vli89-bug/
> https://www.google.com/search?q=intel+c3000+bug
>
> Recurring issue - similar problems were discussed noisily industry wide:
>
> https://en.wikipedia.org/wiki/Silvermont#Erratum
> https://www.servethehome.com/intel-atom-c2000-series-bug-quiet/
> https://www.servethehome.com/intel-atom-c2000-c0-stepping-fixing-the-avr54-bug/
> https://www.google.com/search?q=intel+c2000+bug
>
> As with the previous generation processors and main boards, Intel FAILS.
> Wait for updated CPU revision boards, consult techs and prepare for RMA.
>
> Kind regards,
> Anton Lazarov
Reply | Threaded
Open this post in threaded view
|

Re: SuperMicro A2SDi-4C-HLN4F

lists-2
Fri, 17 Aug 2018 11:15:32 +0000 Rupert Gallagher <[hidden email]>
> FUD
>
> Sent from total and utter ignorance.

Rupert,

Plus the dmesg you posted is by the RAM disk installation OpenBSD kernel.
You Should complete an actual OpenBSD installation to get the full dmesg.

> On Fri, Aug 17, 2018 at 12:24, <[hidden email]> wrote:
>
> > Fri, 17 Aug 2018 06:23:23 +0000 Rupert Gallagher <[hidden email]>  
> >> Using the serial over lan console, the keyboard keeps working, and it is possible to enter the shell. Using the standard console, the keyboard does not work. There is plenty of drivers that need to be written. No, I do not write/port drivers.
> >>
> >> By comparison, both freebsd and centos linux just work out of the box. Both dmesgs are attached as reference.
> >>
> >> The board is really nice.
> >>
> >> On Thu, Aug 16, 2018 at 18:04, Rupert Gallagher <[hidden email]> wrote:
> >>  
> >> > https://www.supermicro.com/products/motherboard/atom/A2SDi-4C-HLN4F.cfm 
> >
> > Hi misc@,
> >
> > This one is much quieter, the vendors have reacted to the previous flaw.
> > As nice as it can get.. before trusting non-technical reviews, be aware:
> >
> > https://en.wikipedia.org/wiki/Goldmont#Erratum
> > https://www.servethehome.com/another-atom-bomb-intel-e3800-bay-trail-atom-vli89-bug/
> > https://www.google.com/search?q=intel+c3000+bug
> >
> > Recurring issue - similar problems were discussed noisily industry wide:
> >
> > https://en.wikipedia.org/wiki/Silvermont#Erratum
> > https://www.servethehome.com/intel-atom-c2000-series-bug-quiet/
> > https://www.servethehome.com/intel-atom-c2000-c0-stepping-fixing-the-avr54-bug/
> > https://www.google.com/search?q=intel+c2000+bug
> >
> > As with the previous generation processors and main boards, Intel FAILS.
> > Wait for updated CPU revision boards, consult techs and prepare for RMA.
> >
> > Kind regards,
> > Anton Lazarov

   "Similar to previous Silvermont generation design flaws were found in
 processor circuitry resulting in cease of operation when processors are
 actively used for several years.  Errata named

 APL47  System May Experience Inability to Boot or May Cease Operation

 and

 APL48  System May Experience Non-functioning GPIO Weak Pull-up Circuitry

 were added to documentation in June 2017 stating:

 The Low Pin Count (LPC), Real-Time Clock (RTC), Secure Digital (SD) card
 and General-Purpose Input/Output (GPIO) interfaces may stop functioning.

 The documents cited in the Wikipedia article, see for current versions..
 that apply to the later CPU stepping models, if any, as time progresses:

 https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/pentium-celeron-n-series-j-series-datasheet-spec-update.pdf
 https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/atom-c3000-family-spec-update.pdf

 APL47 Problem:  Under certain conditions LPC, SD card and RTC circuitry
 may stop functioning in the outer years of use.

 APL48 Problem:  When platform drives the GPIO pin low, GPIOs programmed
 with weak pull-up circuitry may not maintain a value above VIH when not
 actively driven in outer years of service."

 Some other "interesting" errata:

 APL23  HD Audio Recording May Experience a Glitch While Opening or
        Closing Audio Streams

 APL25  USB Device Controller Incorrectly Interprets U3 Wakeup For Warm
        Reset

 APL27  USB 2.0 Timing Responsiveness Degradation

 APL50  System May Unexpectedly Shut Down When Software Requests a Reset

 APL53  Warm Reset May Result in System Hang or Unexpected System
        Behaviour

 Keep your eyes open and.. expect some more technical Atom CPU findings.

Reply | Threaded
Open this post in threaded view
|

Re: SuperMicro A2SDi-4C-HLN4F

Rupert Gallagher
This is a really nice board at a really nice price, and you should stop scaring people off. These Atoms are cleaner than both Cores and Xeons, and AMDs have have their fair share of problems.  Citin g C2000 against the latest C3000 is  FUD. You should come clean first, and cite a comparable server board at a comparable cost and clear of bugs.

On Fri, Aug 17, 2018 at 17:38, <[hidden email]> wrote:

> Fri, 17 Aug 2018 11:15:32 +0000 Rupert Gallagher <[hidden email]>
>> FUD
>>
>> Sent from total and utter ignorance.
>
> Rupert,
>
> Plus the dmesg you posted is by the RAM disk installation OpenBSD kernel.
> You Should complete an actual OpenBSD installation to get the full dmesg.
>
>> On Fri, Aug 17, 2018 at 12:24, <[hidden email]> wrote:
>>
>> > Fri, 17 Aug 2018 06:23:23 +0000 Rupert Gallagher <[hidden email]>
>> >> Using the serial over lan console, the keyboard keeps working, and it is possible to enter the shell. Using the standard console, the keyboard does not work. There is plenty of drivers that need to be written. No, I do not write/port drivers.
>> >>
>> >> By comparison, both freebsd and centos linux just work out of the box. Both dmesgs are attached as reference.
>> >>
>> >> The board is really nice.
>> >>
>> >> On Thu, Aug 16, 2018 at 18:04, Rupert Gallagher <[hidden email]> wrote:
>> >>
>> >> > https://www.supermicro.com/products/motherboard/atom/A2SDi-4C-HLN4F.cfm
>> >
>> > Hi misc@,
>> >
>> > This one is much quieter, the vendors have reacted to the previous flaw.
>> > As nice as it can get.. before trusting non-technical reviews, be aware:
>> >
>> > https://en.wikipedia.org/wiki/Goldmont#Erratum
>> > https://www.servethehome.com/another-atom-bomb-intel-e3800-bay-trail-atom-vli89-bug/
>> > https://www.google.com/search?q=intel+c3000+bug
>> >
>> > Recurring issue - similar problems were discussed noisily industry wide:
>> >
>> > https://en.wikipedia.org/wiki/Silvermont#Erratum
>> > https://www.servethehome.com/intel-atom-c2000-series-bug-quiet/
>> > https://www.servethehome.com/intel-atom-c2000-c0-stepping-fixing-the-avr54-bug/
>> > https://www.google.com/search?q=intel+c2000+bug
>> >
>> > As with the previous generation processors and main boards, Intel FAILS.
>> > Wait for updated CPU revision boards, consult techs and prepare for RMA.
>> >
>> > Kind regards,
>> > Anton Lazarov
>
> "Similar to previous Silvermont generation design flaws were found in
> processor circuitry resulting in cease of operation when processors are
> actively used for several years. Errata named
>
> APL47 System May Experience Inability to Boot or May Cease Operation
>
> and
>
> APL48 System May Experience Non-functioning GPIO Weak Pull-up Circuitry
>
> were added to documentation in June 2017 stating:
>
> The Low Pin Count (LPC), Real-Time Clock (RTC), Secure Digital (SD) card
> and General-Purpose Input/Output (GPIO) interfaces may stop functioning.
>
> The documents cited in the Wikipedia article, see for current versions..
> that apply to the later CPU stepping models, if any, as time progresses:
>
> https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/pentium-celeron-n-series-j-series-datasheet-spec-update.pdf
> https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/atom-c3000-family-spec-update.pdf
>
> APL47 Problem: Under certain conditions LPC, SD card and RTC circuitry
> may stop functioning in the outer years of use.
>
> APL48 Problem: When platform drives the GPIO pin low, GPIOs programmed
> with weak pull-up circuitry may not maintain a value above VIH when not
> actively driven in outer years of service."
>
> Some other "interesting" errata:
>
> APL23 HD Audio Recording May Experience a Glitch While Opening or
> Closing Audio Streams
>
> APL25 USB Device Controller Incorrectly Interprets U3 Wakeup For Warm
> Reset
>
> APL27 USB 2.0 Timing Responsiveness Degradation
>
> APL50 System May Unexpectedly Shut Down When Software Requests a Reset
>
> APL53 Warm Reset May Result in System Hang or Unexpected System
> Behaviour
>
> Keep your eyes open and.. expect some more technical Atom CPU findings.
Reply | Threaded
Open this post in threaded view
|

Re: SuperMicro A2SDi-4C-HLN4F

lists-2
Fri, 17 Aug 2018 16:48:52 +0000 Rupert Gallagher <[hidden email]>
> This is a really nice board at a really nice price, and you should
> stop scaring people off. These Atoms are cleaner than both Cores and
> Xeons, and AMDs have have their fair share of problems.  Citin g
> C2000 against the latest C3000 is  FUD. You should come clean first,
> and cite a comparable server board at a comparable cost and clear of
> bugs.

Rupert,

The publication and processor specifications are about the C3000 CPU bug.
Similar hardware flaws affect the new C3000 CPUs, same microarchitecture.
These Atoms are also fully affected by Intel HW flaws Spectre & Meltdown.

With NONE real hardware understanding, you have to rely upon vendor spec.
With NONE single dmesg after installation, you can NOT advise this board.
With NONE real OpenBSD experience, you better ask the developers instead.

You are speculatively advertising yet unsupported boards as feature sets.
Not interested reading anything that pops in your visibility on the list.
It takes very little effort to have some opinion and I do NOT want yours.

Kind regards,
Anton Lazarov

> On Fri, Aug 17, 2018 at 17:38, <[hidden email]> wrote:
>
> > Fri, 17 Aug 2018 11:15:32 +0000 Rupert Gallagher
> > <[hidden email]>  
> >> FUD
> >>
> >> Sent from total and utter ignorance.  
> >
> > Rupert,
> >
> > Plus the dmesg you posted is by the RAM disk installation OpenBSD
> > kernel. You Should complete an actual OpenBSD installation to get
> > the full dmesg.
> >> On Fri, Aug 17, 2018 at 12:24, <[hidden email]> wrote:
> >>  
> >> > Fri, 17 Aug 2018 06:23:23 +0000 Rupert Gallagher
> >> > <[hidden email]>  
> >> >> Using the serial over lan console, the keyboard keeps working,
> >> >> and it is possible to enter the shell. Using the standard
> >> >> console, the keyboard does not work. There is plenty of drivers
> >> >> that need to be written. No, I do not write/port drivers.
> >> >>
> >> >> By comparison, both freebsd and centos linux just work out of
> >> >> the box. Both dmesgs are attached as reference.
> >> >>
> >> >> The board is really nice.
> >> >>
> >> >> On Thu, Aug 16, 2018 at 18:04, Rupert Gallagher
> >> >> <[hidden email]> wrote:
> >> >> > https://www.supermicro.com/products/motherboard/atom/A2SDi-4C-HLN4F.cfm 
> >> >
> >> > Hi misc@,
> >> >
> >> > This one is much quieter, the vendors have reacted to the
> >> > previous flaw. As nice as it can get.. before trusting
> >> > non-technical reviews, be aware:
> >> >
> >> > https://en.wikipedia.org/wiki/Goldmont#Erratum
> >> > https://www.servethehome.com/another-atom-bomb-intel-e3800-bay-trail-atom-vli89-bug/
> >> > https://www.google.com/search?q=intel+c3000+bug
> >> >
> >> > Recurring issue - similar problems were discussed noisily
> >> > industry wide:
> >> >
> >> > https://en.wikipedia.org/wiki/Silvermont#Erratum
> >> > https://www.servethehome.com/intel-atom-c2000-series-bug-quiet/
> >> > https://www.servethehome.com/intel-atom-c2000-c0-stepping-fixing-the-avr54-bug/
> >> > https://www.google.com/search?q=intel+c2000+bug
> >> >
> >> > As with the previous generation processors and main boards,
> >> > Intel FAILS. Wait for updated CPU revision boards, consult techs
> >> > and prepare for RMA.
> >> >
> >> > Kind regards,
> >> > Anton Lazarov  
> >
> > "Similar to previous Silvermont generation design flaws were found
> > in processor circuitry resulting in cease of operation when
> > processors are actively used for several years. Errata named
> >
> > APL47 System May Experience Inability to Boot or May Cease Operation
> >
> > and
> >
> > APL48 System May Experience Non-functioning GPIO Weak Pull-up
> > Circuitry
> >
> > were added to documentation in June 2017 stating:
> >
> > The Low Pin Count (LPC), Real-Time Clock (RTC), Secure Digital (SD)
> > card and General-Purpose Input/Output (GPIO) interfaces may stop
> > functioning.
> >
> > The documents cited in the Wikipedia article, see for current
> > versions.. that apply to the later CPU stepping models, if any, as
> > time progresses:
> >
> > https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/pentium-celeron-n-series-j-series-datasheet-spec-update.pdf
> > https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/atom-c3000-family-spec-update.pdf
> >
> > APL47 Problem: Under certain conditions LPC, SD card and RTC
> > circuitry may stop functioning in the outer years of use.
> >
> > APL48 Problem: When platform drives the GPIO pin low, GPIOs
> > programmed with weak pull-up circuitry may not maintain a value
> > above VIH when not actively driven in outer years of service."
> >
> > Some other "interesting" errata:
> >
> > APL23 HD Audio Recording May Experience a Glitch While Opening or
> > Closing Audio Streams
> >
> > APL25 USB Device Controller Incorrectly Interprets U3 Wakeup For
> > Warm Reset
> >
> > APL27 USB 2.0 Timing Responsiveness Degradation
> >
> > APL50 System May Unexpectedly Shut Down When Software Requests a
> > Reset
> >
> > APL53 Warm Reset May Result in System Hang or Unexpected System
> > Behaviour
> >
> > Keep your eyes open and.. expect some more technical Atom CPU
> > findings.  

Reply | Threaded
Open this post in threaded view
|

Re: SuperMicro A2SDi-4C-HLN4F

Rupert Gallagher
I can pay you with your same currency.

You have no experience installing openbsd on a new architecture:

> Available disks are: none

Your references are not specific to the C3000 architecture.

Orher architectures are explicitly known for being bugged, but this did not stop the developers.

On Fri, Aug 17, 2018 at 22:09, <[hidden email]> wrote:

> Fri, 17 Aug 2018 16:48:52 +0000 Rupert Gallagher <[hidden email]>
>> This is a really nice board at a really nice price, and you should
>> stop scaring people off. These Atoms are cleaner than both Cores and
>> Xeons, and AMDs have have their fair share of problems. Citin g
>> C2000 against the latest C3000 is FUD. You should come clean first,
>> and cite a comparable server board at a comparable cost and clear of
>> bugs.
>
> Rupert,
>
> The publication and processor specifications are about the C3000 CPU bug.
> Similar hardware flaws affect the new C3000 CPUs, same microarchitecture.
> These Atoms are also fully affected by Intel HW flaws Spectre & Meltdown.
>
> With NONE real hardware understanding, you have to rely upon vendor spec.
> With NONE single dmesg after installation, you can NOT advise this board.
> With NONE real OpenBSD experience, you better ask the developers instead.
>
> You are speculatively advertising yet unsupported boards as feature sets.
> Not interested reading anything that pops in your visibility on the list.
> It takes very little effort to have some opinion and I do NOT want yours.
>
> Kind regards,
> Anton Lazarov
>
>> On Fri, Aug 17, 2018 at 17:38, <[hidden email]> wrote:
>>
>> > Fri, 17 Aug 2018 11:15:32 +0000 Rupert Gallagher
>> > <[hidden email]>
>> >> FUD
>> >>
>> >> Sent from total and utter ignorance.
>> >
>> > Rupert,
>> >
>> > Plus the dmesg you posted is by the RAM disk installation OpenBSD
>> > kernel. You Should complete an actual OpenBSD installation to get
>> > the full dmesg.
>> >> On Fri, Aug 17, 2018 at 12:24, <[hidden email]> wrote:
>> >>
>> >> > Fri, 17 Aug 2018 06:23:23 +0000 Rupert Gallagher
>> >> > <[hidden email]>
>> >> >> Using the serial over lan console, the keyboard keeps working,
>> >> >> and it is possible to enter the shell. Using the standard
>> >> >> console, the keyboard does not work. There is plenty of drivers
>> >> >> that need to be written. No, I do not write/port drivers.
>> >> >>
>> >> >> By comparison, both freebsd and centos linux just work out of
>> >> >> the box. Both dmesgs are attached as reference.
>> >> >>
>> >> >> The board is really nice.
>> >> >>
>> >> >> On Thu, Aug 16, 2018 at 18:04, Rupert Gallagher
>> >> >> <[hidden email]> wrote:
>> >> >> > https://www.supermicro.com/products/motherboard/atom/A2SDi-4C-HLN4F.cfm
>> >> >
>> >> > Hi misc@,
>> >> >
>> >> > This one is much quieter, the vendors have reacted to the
>> >> > previous flaw. As nice as it can get.. before trusting
>> >> > non-technical reviews, be aware:
>> >> >
>> >> > https://en.wikipedia.org/wiki/Goldmont#Erratum
>> >> > https://www.servethehome.com/another-atom-bomb-intel-e3800-bay-trail-atom-vli89-bug/
>> >> > https://www.google.com/search?q=intel+c3000+bug
>> >> >
>> >> > Recurring issue - similar problems were discussed noisily
>> >> > industry wide:
>> >> >
>> >> > https://en.wikipedia.org/wiki/Silvermont#Erratum
>> >> > https://www.servethehome.com/intel-atom-c2000-series-bug-quiet/
>> >> > https://www.servethehome.com/intel-atom-c2000-c0-stepping-fixing-the-avr54-bug/
>> >> > https://www.google.com/search?q=intel+c2000+bug
>> >> >
>> >> > As with the previous generation processors and main boards,
>> >> > Intel FAILS. Wait for updated CPU revision boards, consult techs
>> >> > and prepare for RMA.
>> >> >
>> >> > Kind regards,
>> >> > Anton Lazarov
>> >
>> > "Similar to previous Silvermont generation design flaws were found
>> > in processor circuitry resulting in cease of operation when
>> > processors are actively used for several years. Errata named
>> >
>> > APL47 System May Experience Inability to Boot or May Cease Operation
>> >
>> > and
>> >
>> > APL48 System May Experience Non-functioning GPIO Weak Pull-up
>> > Circuitry
>> >
>> > were added to documentation in June 2017 stating:
>> >
>> > The Low Pin Count (LPC), Real-Time Clock (RTC), Secure Digital (SD)
>> > card and General-Purpose Input/Output (GPIO) interfaces may stop
>> > functioning.
>> >
>> > The documents cited in the Wikipedia article, see for current
>> > versions.. that apply to the later CPU stepping models, if any, as
>> > time progresses:
>> >
>> > https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/pentium-celeron-n-series-j-series-datasheet-spec-update.pdf
>> > https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/atom-c3000-family-spec-update.pdf
>> >
>> > APL47 Problem: Under certain conditions LPC, SD card and RTC
>> > circuitry may stop functioning in the outer years of use.
>> >
>> > APL48 Problem: When platform drives the GPIO pin low, GPIOs
>> > programmed with weak pull-up circuitry may not maintain a value
>> > above VIH when not actively driven in outer years of service."
>> >
>> > Some other "interesting" errata:
>> >
>> > APL23 HD Audio Recording May Experience a Glitch While Opening or
>> > Closing Audio Streams
>> >
>> > APL25 USB Device Controller Incorrectly Interprets U3 Wakeup For
>> > Warm Reset
>> >
>> > APL27 USB 2.0 Timing Responsiveness Degradation
>> >
>> > APL50 System May Unexpectedly Shut Down When Software Requests a
>> > Reset
>> >
>> > APL53 Warm Reset May Result in System Hang or Unexpected System
>> > Behaviour
>> >
>> > Keep your eyes open and.. expect some more technical Atom CPU
>> > findings.
Reply | Threaded
Open this post in threaded view
|

Re: SuperMicro A2SDi-4C-HLN4F

lists-2
Sat, 18 Aug 2018 08:29:33 +0000 Rupert Gallagher <[hidden email]>
> Orher architectures are explicitly known for being bugged, but this
> did not stop the developers.

Hi Rupert,

Enough already, just post a full dmesg when you can.  The hardware is not
perfect, watch out for unfixed CPU bugs as these are soldered down chips.

They also include a lot more than the CPU where most of the problems are.
So keep an eye on these boards and be a bit more impartial and objective.

It would be so nice if you could ship new boards like these to developers
or donate some of your excess hardware costs so that they get the boards.

That might be much more helpful than complaints over the installer dmesg.

Kind regards,
Anton Lazarov

> On Fri, Aug 17, 2018 at 22:09, <[hidden email]> wrote:
>
> > Fri, 17 Aug 2018 16:48:52 +0000 Rupert Gallagher <[hidden email]>  
> >> This is a really nice board at a really nice price, and you should
> >> stop scaring people off. These Atoms are cleaner than both Cores and
> >> Xeons, and AMDs have have their fair share of problems. Citin g
> >> C2000 against the latest C3000 is FUD. You should come clean first,
> >> and cite a comparable server board at a comparable cost and clear of
> >> bugs.  
> >
> > Rupert,
> >
> > The publication and processor specifications are about the C3000 CPU bug.
> > Similar hardware flaws affect the new C3000 CPUs, same microarchitecture.
> > These Atoms are also fully affected by Intel HW flaws Spectre & Meltdown.
> >
> > With NONE real hardware understanding, you have to rely upon vendor spec.
> > With NONE single dmesg after installation, you can NOT advise this board.
> > With NONE real OpenBSD experience, you better ask the developers instead.
> >
> > You are speculatively advertising yet unsupported boards as feature sets.
> > Not interested reading anything that pops in your visibility on the list.
> > It takes very little effort to have some opinion and I do NOT want yours.
> >
> > Kind regards,
> > Anton Lazarov
> >  
> >> On Fri, Aug 17, 2018 at 17:38, <[hidden email]> wrote:
> >>  
> >> > Fri, 17 Aug 2018 11:15:32 +0000 Rupert Gallagher
> >> > <[hidden email]>  
> >> >> FUD
> >> >>
> >> >> Sent from total and utter ignorance.  
> >> >
> >> > Rupert,
> >> >
> >> > Plus the dmesg you posted is by the RAM disk installation OpenBSD
> >> > kernel. You Should complete an actual OpenBSD installation to get
> >> > the full dmesg.  
> >> >> On Fri, Aug 17, 2018 at 12:24, <[hidden email]> wrote:
> >> >>  
> >> >> > Fri, 17 Aug 2018 06:23:23 +0000 Rupert Gallagher
> >> >> > <[hidden email]>  
> >> >> >> Using the serial over lan console, the keyboard keeps working,
> >> >> >> and it is possible to enter the shell. Using the standard
> >> >> >> console, the keyboard does not work. There is plenty of drivers
> >> >> >> that need to be written. No, I do not write/port drivers.
> >> >> >>
> >> >> >> By comparison, both freebsd and centos linux just work out of
> >> >> >> the box. Both dmesgs are attached as reference.
> >> >> >>
> >> >> >> The board is really nice.
> >> >> >>
> >> >> >> On Thu, Aug 16, 2018 at 18:04, Rupert Gallagher
> >> >> >> <[hidden email]> wrote:  
> >> >> >> > https://www.supermicro.com/products/motherboard/atom/A2SDi-4C-HLN4F.cfm 
> >> >> >
> >> >> > Hi misc@,
> >> >> >
> >> >> > This one is much quieter, the vendors have reacted to the
> >> >> > previous flaw. As nice as it can get.. before trusting
> >> >> > non-technical reviews, be aware:
> >> >> >
> >> >> > https://en.wikipedia.org/wiki/Goldmont#Erratum
> >> >> > https://www.servethehome.com/another-atom-bomb-intel-e3800-bay-trail-atom-vli89-bug/
> >> >> > https://www.google.com/search?q=intel+c3000+bug
> >> >> >
> >> >> > Recurring issue - similar problems were discussed noisily
> >> >> > industry wide:
> >> >> >
> >> >> > https://en.wikipedia.org/wiki/Silvermont#Erratum
> >> >> > https://www.servethehome.com/intel-atom-c2000-series-bug-quiet/
> >> >> > https://www.servethehome.com/intel-atom-c2000-c0-stepping-fixing-the-avr54-bug/
> >> >> > https://www.google.com/search?q=intel+c2000+bug
> >> >> >
> >> >> > As with the previous generation processors and main boards,
> >> >> > Intel FAILS. Wait for updated CPU revision boards, consult techs
> >> >> > and prepare for RMA.
> >> >> >
> >> >> > Kind regards,
> >> >> > Anton Lazarov  
> >> >
> >> > "Similar to previous Silvermont generation design flaws were found
> >> > in processor circuitry resulting in cease of operation when
> >> > processors are actively used for several years. Errata named
> >> >
> >> > APL47 System May Experience Inability to Boot or May Cease Operation
> >> >
> >> > and
> >> >
> >> > APL48 System May Experience Non-functioning GPIO Weak Pull-up
> >> > Circuitry
> >> >
> >> > were added to documentation in June 2017 stating:
> >> >
> >> > The Low Pin Count (LPC), Real-Time Clock (RTC), Secure Digital (SD)
> >> > card and General-Purpose Input/Output (GPIO) interfaces may stop
> >> > functioning.
> >> >
> >> > The documents cited in the Wikipedia article, see for current
> >> > versions.. that apply to the later CPU stepping models, if any, as
> >> > time progresses:
> >> >
> >> > https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/pentium-celeron-n-series-j-series-datasheet-spec-update.pdf
> >> > https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/atom-c3000-family-spec-update.pdf
> >> >
> >> > APL47 Problem: Under certain conditions LPC, SD card and RTC
> >> > circuitry may stop functioning in the outer years of use.
> >> >
> >> > APL48 Problem: When platform drives the GPIO pin low, GPIOs
> >> > programmed with weak pull-up circuitry may not maintain a value
> >> > above VIH when not actively driven in outer years of service."
> >> >
> >> > Some other "interesting" errata:
> >> >
> >> > APL23 HD Audio Recording May Experience a Glitch While Opening or
> >> > Closing Audio Streams
> >> >
> >> > APL25 USB Device Controller Incorrectly Interprets U3 Wakeup For
> >> > Warm Reset
> >> >
> >> > APL27 USB 2.0 Timing Responsiveness Degradation
> >> >
> >> > APL50 System May Unexpectedly Shut Down When Software Requests a
> >> > Reset
> >> >
> >> > APL53 Warm Reset May Result in System Hang or Unexpected System
> >> > Behaviour
> >> >
> >> > Keep your eyes open and.. expect some more technical Atom CPU
> >> > findings  

Reply | Threaded
Open this post in threaded view
|

Re: SuperMicro A2SDi-4C-HLN4F

Rupert Gallagher
I am not complaining about obsd.

I am complaining about your FUD on C3000 and your request for a post-installation dmesg. No installation is possible for lack of disk drivers. I do not develop/port drivers. The dmesgs I already posted are all I can share.

The only way to run obsd at this time is via qemu/proxmox or vmware/esxi.

On Sat, Aug 18, 2018 at 12:55, <[hidden email]> wrote:

> Sat, 18 Aug 2018 08:29:33 +0000 Rupert Gallagher <[hidden email]>
>> Orher architectures are explicitly known for being bugged, but this
>> did not stop the developers.
>
> Hi Rupert,
>
> Enough already, just post a full dmesg when you can. The hardware is not
> perfect, watch out for unfixed CPU bugs as these are soldered down chips.
>
> They also include a lot more than the CPU where most of the problems are.
> So keep an eye on these boards and be a bit more impartial and objective.
>
> It would be so nice if you could ship new boards like these to developers
> or donate some of your excess hardware costs so that they get the boards.
>
> That might be much more helpful than complaints over the installer dmesg.
>
> Kind regards,
> Anton Lazarov
>
>> On Fri, Aug 17, 2018 at 22:09, <[hidden email]> wrote:
>>
>> > Fri, 17 Aug 2018 16:48:52 +0000 Rupert Gallagher <[hidden email]>
>> >> This is a really nice board at a really nice price, and you should
>> >> stop scaring people off. These Atoms are cleaner than both Cores and
>> >> Xeons, and AMDs have have their fair share of problems. Citin g
>> >> C2000 against the latest C3000 is FUD. You should come clean first,
>> >> and cite a comparable server board at a comparable cost and clear of
>> >> bugs.
>> >
>> > Rupert,
>> >
>> > The publication and processor specifications are about the C3000 CPU bug.
>> > Similar hardware flaws affect the new C3000 CPUs, same microarchitecture.
>> > These Atoms are also fully affected by Intel HW flaws Spectre & Meltdown.
>> >
>> > With NONE real hardware understanding, you have to rely upon vendor spec.
>> > With NONE single dmesg after installation, you can NOT advise this board.
>> > With NONE real OpenBSD experience, you better ask the developers instead.
>> >
>> > You are speculatively advertising yet unsupported boards as feature sets.
>> > Not interested reading anything that pops in your visibility on the list.
>> > It takes very little effort to have some opinion and I do NOT want yours.
>> >
>> > Kind regards,
>> > Anton Lazarov
>> >
>> >> On Fri, Aug 17, 2018 at 17:38, <[hidden email]> wrote:
>> >>
>> >> > Fri, 17 Aug 2018 11:15:32 +0000 Rupert Gallagher
>> >> > <[hidden email]>
>> >> >> FUD
>> >> >>
>> >> >> Sent from total and utter ignorance.
>> >> >
>> >> > Rupert,
>> >> >
>> >> > Plus the dmesg you posted is by the RAM disk installation OpenBSD
>> >> > kernel. You Should complete an actual OpenBSD installation to get
>> >> > the full dmesg.
>> >> >> On Fri, Aug 17, 2018 at 12:24, <[hidden email]> wrote:
>> >> >>
>> >> >> > Fri, 17 Aug 2018 06:23:23 +0000 Rupert Gallagher
>> >> >> > <[hidden email]>
>> >> >> >> Using the serial over lan console, the keyboard keeps working,
>> >> >> >> and it is possible to enter the shell. Using the standard
>> >> >> >> console, the keyboard does not work. There is plenty of drivers
>> >> >> >> that need to be written. No, I do not write/port drivers.
>> >> >> >>
>> >> >> >> By comparison, both freebsd and centos linux just work out of
>> >> >> >> the box. Both dmesgs are attached as reference.
>> >> >> >>
>> >> >> >> The board is really nice.
>> >> >> >>
>> >> >> >> On Thu, Aug 16, 2018 at 18:04, Rupert Gallagher
>> >> >> >> <[hidden email]> wrote:
>> >> >> >> > https://www.supermicro.com/products/motherboard/atom/A2SDi-4C-HLN4F.cfm
>> >> >> >
>> >> >> > Hi misc@,
>> >> >> >
>> >> >> > This one is much quieter, the vendors have reacted to the
>> >> >> > previous flaw. As nice as it can get.. before trusting
>> >> >> > non-technical reviews, be aware:
>> >> >> >
>> >> >> > https://en.wikipedia.org/wiki/Goldmont#Erratum
>> >> >> > https://www.servethehome.com/another-atom-bomb-intel-e3800-bay-trail-atom-vli89-bug/
>> >> >> > https://www.google.com/search?q=intel+c3000+bug
>> >> >> >
>> >> >> > Recurring issue - similar problems were discussed noisily
>> >> >> > industry wide:
>> >> >> >
>> >> >> > https://en.wikipedia.org/wiki/Silvermont#Erratum
>> >> >> > https://www.servethehome.com/intel-atom-c2000-series-bug-quiet/
>> >> >> > https://www.servethehome.com/intel-atom-c2000-c0-stepping-fixing-the-avr54-bug/
>> >> >> > https://www.google.com/search?q=intel+c2000+bug
>> >> >> >
>> >> >> > As with the previous generation processors and main boards,
>> >> >> > Intel FAILS. Wait for updated CPU revision boards, consult techs
>> >> >> > and prepare for RMA.
>> >> >> >
>> >> >> > Kind regards,
>> >> >> > Anton Lazarov
>> >> >
>> >> > "Similar to previous Silvermont generation design flaws were found
>> >> > in processor circuitry resulting in cease of operation when
>> >> > processors are actively used for several years. Errata named
>> >> >
>> >> > APL47 System May Experience Inability to Boot or May Cease Operation
>> >> >
>> >> > and
>> >> >
>> >> > APL48 System May Experience Non-functioning GPIO Weak Pull-up
>> >> > Circuitry
>> >> >
>> >> > were added to documentation in June 2017 stating:
>> >> >
>> >> > The Low Pin Count (LPC), Real-Time Clock (RTC), Secure Digital (SD)
>> >> > card and General-Purpose Input/Output (GPIO) interfaces may stop
>> >> > functioning.
>> >> >
>> >> > The documents cited in the Wikipedia article, see for current
>> >> > versions.. that apply to the later CPU stepping models, if any, as
>> >> > time progresses:
>> >> >
>> >> > https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/pentium-celeron-n-series-j-series-datasheet-spec-update.pdf
>> >> > https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/atom-c3000-family-spec-update.pdf
>> >> >
>> >> > APL47 Problem: Under certain conditions LPC, SD card and RTC
>> >> > circuitry may stop functioning in the outer years of use.
>> >> >
>> >> > APL48 Problem: When platform drives the GPIO pin low, GPIOs
>> >> > programmed with weak pull-up circuitry may not maintain a value
>> >> > above VIH when not actively driven in outer years of service."
>> >> >
>> >> > Some other "interesting" errata:
>> >> >
>> >> > APL23 HD Audio Recording May Experience a Glitch While Opening or
>> >> > Closing Audio Streams
>> >> >
>> >> > APL25 USB Device Controller Incorrectly Interprets U3 Wakeup For
>> >> > Warm Reset
>> >> >
>> >> > APL27 USB 2.0 Timing Responsiveness Degradation
>> >> >
>> >> > APL50 System May Unexpectedly Shut Down When Software Requests a
>> >> > Reset
>> >> >
>> >> > APL53 Warm Reset May Result in System Hang or Unexpected System
>> >> > Behaviour
>> >> >
>> >> > Keep your eyes open and.. expect some more technical Atom CPU
>> >> > findings
Reply | Threaded
Open this post in threaded view
|

Re: SuperMicro A2SDi-4C-HLN4F

lists-2
Sat, 18 Aug 2018 12:28:17 +0000 Rupert Gallagher <[hidden email]>
> I am not complaining about obsd.
>
> I am complaining about your FUD on C3000 and your request for a
> post-installation dmesg. No installation is possible for lack of disk
> drivers. I do not develop/port drivers. The dmesgs I already posted
> are all I can share.
>
> The only way to run obsd at this time is via qemu/proxmox or
> vmware/esxi.

Rupert,

You would not have to complain if you hadn't spammed your puny pretences.
The boards and processors have issues to be known, accompanying warnings.

The dmesg you posted does not help anything.  It looks like a poorly done
commercial for other operating systems that "supposedly" work.  Only, you
have not been running anything past boot up.  Accept it, you are a troll.

Looks like you point out there are "drivers to be written", for your use.
I call this helpless whining.  I told you how you can help this process..

Get the boards in the hands of developers with OpenBSD who write drivers.
And first kindly ask about it whether it is possible to fit in busy work.

Kind regards,
Anton Lazarov

> On Sat, Aug 18, 2018 at 12:55, <[hidden email]> wrote:
>
> > Sat, 18 Aug 2018 08:29:33 +0000 Rupert Gallagher <[hidden email]>  
> >> Orher architectures are explicitly known for being bugged, but this
> >> did not stop the developers.  
> >
> > Hi Rupert,
> >
> > Enough already, just post a full dmesg when you can. The hardware is not
> > perfect, watch out for unfixed CPU bugs as these are soldered down chips.
> >
> > They also include a lot more than the CPU where most of the problems are.
> > So keep an eye on these boards and be a bit more impartial and objective.
> >
> > It would be so nice if you could ship new boards like these to developers
> > or donate some of your excess hardware costs so that they get the boards.
> >
> > That might be much more helpful than complaints over the installer dmesg.
> >
> > Kind regards,
> > Anton Lazarov
> >  
> >> On Fri, Aug 17, 2018 at 22:09, <[hidden email]> wrote:
> >>  
> >> > Fri, 17 Aug 2018 16:48:52 +0000 Rupert Gallagher <[hidden email]>  
> >> >> This is a really nice board at a really nice price, and you should
> >> >> stop scaring people off. These Atoms are cleaner than both Cores and
> >> >> Xeons, and AMDs have have their fair share of problems. Citin g
> >> >> C2000 against the latest C3000 is FUD. You should come clean first,
> >> >> and cite a comparable server board at a comparable cost and clear of
> >> >> bugs.  
> >> >
> >> > Rupert,
> >> >
> >> > The publication and processor specifications are about the C3000 CPU bug.
> >> > Similar hardware flaws affect the new C3000 CPUs, same microarchitecture.
> >> > These Atoms are also fully affected by Intel HW flaws Spectre & Meltdown.
> >> >
> >> > With NONE real hardware understanding, you have to rely upon vendor spec.
> >> > With NONE single dmesg after installation, you can NOT advise this board.
> >> > With NONE real OpenBSD experience, you better ask the developers instead.
> >> >
> >> > You are speculatively advertising yet unsupported boards as feature sets.
> >> > Not interested reading anything that pops in your visibility on the list.
> >> > It takes very little effort to have some opinion and I do NOT want yours.
> >> >
> >> > Kind regards,
> >> > Anton Lazarov
> >> >  
> >> >> On Fri, Aug 17, 2018 at 17:38, <[hidden email]> wrote:
> >> >>  
> >> >> > Fri, 17 Aug 2018 11:15:32 +0000 Rupert Gallagher
> >> >> > <[hidden email]>  
> >> >> >> FUD
> >> >> >>
> >> >> >> Sent from total and utter ignorance.  
> >> >> >
> >> >> > Rupert,
> >> >> >
> >> >> > Plus the dmesg you posted is by the RAM disk installation OpenBSD
> >> >> > kernel. You Should complete an actual OpenBSD installation to get
> >> >> > the full dmesg.  
> >> >> >> On Fri, Aug 17, 2018 at 12:24, <[hidden email]> wrote:
> >> >> >>  
> >> >> >> > Fri, 17 Aug 2018 06:23:23 +0000 Rupert Gallagher
> >> >> >> > <[hidden email]>  
> >> >> >> >> Using the serial over lan console, the keyboard keeps working,
> >> >> >> >> and it is possible to enter the shell. Using the standard
> >> >> >> >> console, the keyboard does not work. There is plenty of drivers
> >> >> >> >> that need to be written. No, I do not write/port drivers.
> >> >> >> >>
> >> >> >> >> By comparison, both freebsd and centos linux just work out of
> >> >> >> >> the box. Both dmesgs are attached as reference.
> >> >> >> >>
> >> >> >> >> The board is really nice.
> >> >> >> >>
> >> >> >> >> On Thu, Aug 16, 2018 at 18:04, Rupert Gallagher
> >> >> >> >> <[hidden email]> wrote:  
> >> >> >> >> > https://www.supermicro.com/products/motherboard/atom/A2SDi-4C-HLN4F.cfm 
> >> >> >> >
> >> >> >> > Hi misc@,
> >> >> >> >
> >> >> >> > This one is much quieter, the vendors have reacted to the
> >> >> >> > previous flaw. As nice as it can get.. before trusting
> >> >> >> > non-technical reviews, be aware:
> >> >> >> >
> >> >> >> > https://en.wikipedia.org/wiki/Goldmont#Erratum
> >> >> >> > https://www.servethehome.com/another-atom-bomb-intel-e3800-bay-trail-atom-vli89-bug/
> >> >> >> > https://www.google.com/search?q=intel+c3000+bug
> >> >> >> >
> >> >> >> > Recurring issue - similar problems were discussed noisily
> >> >> >> > industry wide:
> >> >> >> >
> >> >> >> > https://en.wikipedia.org/wiki/Silvermont#Erratum
> >> >> >> > https://www.servethehome.com/intel-atom-c2000-series-bug-quiet/
> >> >> >> > https://www.servethehome.com/intel-atom-c2000-c0-stepping-fixing-the-avr54-bug/
> >> >> >> > https://www.google.com/search?q=intel+c2000+bug
> >> >> >> >
> >> >> >> > As with the previous generation processors and main boards,
> >> >> >> > Intel FAILS. Wait for updated CPU revision boards, consult techs
> >> >> >> > and prepare for RMA.
> >> >> >> >
> >> >> >> > Kind regards,
> >> >> >> > Anton Lazarov  
> >> >> >
> >> >> > "Similar to previous Silvermont generation design flaws were found
> >> >> > in processor circuitry resulting in cease of operation when
> >> >> > processors are actively used for several years. Errata named
> >> >> >
> >> >> > APL47 System May Experience Inability to Boot or May Cease Operation
> >> >> >
> >> >> > and
> >> >> >
> >> >> > APL48 System May Experience Non-functioning GPIO Weak Pull-up
> >> >> > Circuitry
> >> >> >
> >> >> > were added to documentation in June 2017 stating:
> >> >> >
> >> >> > The Low Pin Count (LPC), Real-Time Clock (RTC), Secure Digital (SD)
> >> >> > card and General-Purpose Input/Output (GPIO) interfaces may stop
> >> >> > functioning.
> >> >> >
> >> >> > The documents cited in the Wikipedia article, see for current
> >> >> > versions.. that apply to the later CPU stepping models, if any, as
> >> >> > time progresses:
> >> >> >
> >> >> > https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/pentium-celeron-n-series-j-series-datasheet-spec-update.pdf
> >> >> > https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/atom-c3000-family-spec-update.pdf
> >> >> >
> >> >> > APL47 Problem: Under certain conditions LPC, SD card and RTC
> >> >> > circuitry may stop functioning in the outer years of use.
> >> >> >
> >> >> > APL48 Problem: When platform drives the GPIO pin low, GPIOs
> >> >> > programmed with weak pull-up circuitry may not maintain a value
> >> >> > above VIH when not actively driven in outer years of service."
> >> >> >
> >> >> > Some other "interesting" errata:
> >> >> >
> >> >> > APL23 HD Audio Recording May Experience a Glitch While Opening or
> >> >> > Closing Audio Streams
> >> >> >
> >> >> > APL25 USB Device Controller Incorrectly Interprets U3 Wakeup For
> >> >> > Warm Reset
> >> >> >
> >> >> > APL27 USB 2.0 Timing Responsiveness Degradation
> >> >> >
> >> >> > APL50 System May Unexpectedly Shut Down When Software Requests a
> >> >> > Reset
> >> >> >
> >> >> > APL53 Warm Reset May Result in System Hang or Unexpected System
> >> >> > Behaviour
> >> >> >
> >> >> > Keep your eyes open and.. expect some more technical Atom CPU
> >> >> > finding  

Reply | Threaded
Open this post in threaded view
|

Re: SuperMicro A2SDi-4C-HLN4F

Rupert Gallagher
Troll here...

Step out of your nerdiness, and step into  the real world with this scenario. A well-known vendor of server boards has an entite product line based on intel atom C3000 cpus. Each product has a list of certified operating systems, including windows, freebsd, red hat linux enterprise, esxi. The list does not include openbsd. You spend a few days testing the server with various oss. It turns out that the server *works* with everything linux, including debian, proxmox, centos. It also turns out that openbsd fails to recognise the devices, including the disks, and the keyboard. Then you post its dmesg, because obsd folks really like it, together with a positive dmesg from centos. Someone intervenes spreading FUD about the server and flaming you are a troll.

You know what? Fuck you.

Sent from ProtonMail Mobile

On Sun, Aug 19, 2018 at 00:06, <[hidden email]> wrote:

> Sat, 18 Aug 2018 12:28:17 +0000 Rupert Gallagher <[hidden email]>
>> I am not complaining about obsd.
>>
>> I am complaining about your FUD on C3000 and your request for a
>> post-installation dmesg. No installation is possible for lack of disk
>> drivers. I do not develop/port drivers. The dmesgs I already posted
>> are all I can share.
>>
>> The only way to run obsd at this time is via qemu/proxmox or
>> vmware/esxi.
>
> Rupert,
>
> You would not have to complain if you hadn't spammed your puny pretences.
> The boards and processors have issues to be known, accompanying warnings.
>
> The dmesg you posted does not help anything. It looks like a poorly done
> commercial for other operating systems that "supposedly" work. Only, you
> have not been running anything past boot up. Accept it, you are a troll.
>
> Looks like you point out there are "drivers to be written", for your use.
> I call this helpless whining. I told you how you can help this process..
>
> Get the boards in the hands of developers with OpenBSD who write drivers.
> And first kindly ask about it whether it is possible to fit in busy work.
>
> Kind regards,
> Anton Lazarov
>
>> On Sat, Aug 18, 2018 at 12:55, <[hidden email]> wrote:
>>
>> > Sat, 18 Aug 2018 08:29:33 +0000 Rupert Gallagher <[hidden email]>
>> >> Orher architectures are explicitly known for being bugged, but this
>> >> did not stop the developers.
>> >
>> > Hi Rupert,
>> >
>> > Enough already, just post a full dmesg when you can. The hardware is not
>> > perfect, watch out for unfixed CPU bugs as these are soldered down chips.
>> >
>> > They also include a lot more than the CPU where most of the problems are.
>> > So keep an eye on these boards and be a bit more impartial and objective.
>> >
>> > It would be so nice if you could ship new boards like these to developers
>> > or donate some of your excess hardware costs so that they get the boards.
>> >
>> > That might be much more helpful than complaints over the installer dmesg.
>> >
>> > Kind regards,
>> > Anton Lazarov
>> >
>> >> On Fri, Aug 17, 2018 at 22:09, <[hidden email]> wrote:
>> >>
>> >> > Fri, 17 Aug 2018 16:48:52 +0000 Rupert Gallagher <[hidden email]>
>> >> >> This is a really nice board at a really nice price, and you should
>> >> >> stop scaring people off. These Atoms are cleaner than both Cores and
>> >> >> Xeons, and AMDs have have their fair share of problems. Citin g
>> >> >> C2000 against the latest C3000 is FUD. You should come clean first,
>> >> >> and cite a comparable server board at a comparable cost and clear of
>> >> >> bugs.
>> >> >
>> >> > Rupert,
>> >> >
>> >> > The publication and processor specifications are about the C3000 CPU bug.
>> >> > Similar hardware flaws affect the new C3000 CPUs, same microarchitecture.
>> >> > These Atoms are also fully affected by Intel HW flaws Spectre & Meltdown.
>> >> >
>> >> > With NONE real hardware understanding, you have to rely upon vendor spec.
>> >> > With NONE single dmesg after installation, you can NOT advise this board.
>> >> > With NONE real OpenBSD experience, you better ask the developers instead.
>> >> >
>> >> > You are speculatively advertising yet unsupported boards as feature sets.
>> >> > Not interested reading anything that pops in your visibility on the list.
>> >> > It takes very little effort to have some opinion and I do NOT want yours.
>> >> >
>> >> > Kind regards,
>> >> > Anton Lazarov
>> >> >
>> >> >> On Fri, Aug 17, 2018 at 17:38, <[hidden email]> wrote:
>> >> >>
>> >> >> > Fri, 17 Aug 2018 11:15:32 +0000 Rupert Gallagher
>> >> >> > <[hidden email]>
>> >> >> >> FUD
>> >> >> >>
>> >> >> >> Sent from total and utter ignorance.
>> >> >> >
>> >> >> > Rupert,
>> >> >> >
>> >> >> > Plus the dmesg you posted is by the RAM disk installation OpenBSD
>> >> >> > kernel. You Should complete an actual OpenBSD installation to get
>> >> >> > the full dmesg.
>> >> >> >> On Fri, Aug 17, 2018 at 12:24, <[hidden email]> wrote:
>> >> >> >>
>> >> >> >> > Fri, 17 Aug 2018 06:23:23 +0000 Rupert Gallagher
>> >> >> >> > <[hidden email]>
>> >> >> >> >> Using the serial over lan console, the keyboard keeps working,
>> >> >> >> >> and it is possible to enter the shell. Using the standard
>> >> >> >> >> console, the keyboard does not work. There is plenty of drivers
>> >> >> >> >> that need to be written. No, I do not write/port drivers.
>> >> >> >> >>
>> >> >> >> >> By comparison, both freebsd and centos linux just work out of
>> >> >> >> >> the box. Both dmesgs are attached as reference.
>> >> >> >> >>
>> >> >> >> >> The board is really nice.
>> >> >> >> >>
>> >> >> >> >> On Thu, Aug 16, 2018 at 18:04, Rupert Gallagher
>> >> >> >> >> <[hidden email]> wrote:
>> >> >> >> >> > https://www.supermicro.com/products/motherboard/atom/A2SDi-4C-HLN4F.cfm
>> >> >> >> >
>> >> >> >> > Hi misc@,
>> >> >> >> >
>> >> >> >> > This one is much quieter, the vendors have reacted to the
>> >> >> >> > previous flaw. As nice as it can get.. before trusting
>> >> >> >> > non-technical reviews, be aware:
>> >> >> >> >
>> >> >> >> > https://en.wikipedia.org/wiki/Goldmont#Erratum
>> >> >> >> > https://www.servethehome.com/another-atom-bomb-intel-e3800-bay-trail-atom-vli89-bug/
>> >> >> >> > https://www.google.com/search?q=intel+c3000+bug
>> >> >> >> >
>> >> >> >> > Recurring issue - similar problems were discussed noisily
>> >> >> >> > industry wide:
>> >> >> >> >
>> >> >> >> > https://en.wikipedia.org/wiki/Silvermont#Erratum
>> >> >> >> > https://www.servethehome.com/intel-atom-c2000-series-bug-quiet/
>> >> >> >> > https://www.servethehome.com/intel-atom-c2000-c0-stepping-fixing-the-avr54-bug/
>> >> >> >> > https://www.google.com/search?q=intel+c2000+bug
>> >> >> >> >
>> >> >> >> > As with the previous generation processors and main boards,
>> >> >> >> > Intel FAILS. Wait for updated CPU revision boards, consult techs
>> >> >> >> > and prepare for RMA.
>> >> >> >> >
>> >> >> >> > Kind regards,
>> >> >> >> > Anton Lazarov
>> >> >> >
>> >> >> > "Similar to previous Silvermont generation design flaws were found
>> >> >> > in processor circuitry resulting in cease of operation when
>> >> >> > processors are actively used for several years. Errata named
>> >> >> >
>> >> >> > APL47 System May Experience Inability to Boot or May Cease Operation
>> >> >> >
>> >> >> > and
>> >> >> >
>> >> >> > APL48 System May Experience Non-functioning GPIO Weak Pull-up
>> >> >> > Circuitry
>> >> >> >
>> >> >> > were added to documentation in June 2017 stating:
>> >> >> >
>> >> >> > The Low Pin Count (LPC), Real-Time Clock (RTC), Secure Digital (SD)
>> >> >> > card and General-Purpose Input/Output (GPIO) interfaces may stop
>> >> >> > functioning.
>> >> >> >
>> >> >> > The documents cited in the Wikipedia article, see for current
>> >> >> > versions.. that apply to the later CPU stepping models, if any, as
>> >> >> > time progresses:
>> >> >> >
>> >> >> > https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/pentium-celeron-n-series-j-series-datasheet-spec-update.pdf
>> >> >> > https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/atom-c3000-family-spec-update.pdf
>> >> >> >
>> >> >> > APL47 Problem: Under certain conditions LPC, SD card and RTC
>> >> >> > circuitry may stop functioning in the outer years of use.
>> >> >> >
>> >> >> > APL48 Problem: When platform drives the GPIO pin low, GPIOs
>> >> >> > programmed with weak pull-up circuitry may not maintain a value
>> >> >> > above VIH when not actively driven in outer years of service."
>> >> >> >
>> >> >> > Some other "interesting" errata:
>> >> >> >
>> >> >> > APL23 HD Audio Recording May Experience a Glitch While Opening or
>> >> >> > Closing Audio Streams
>> >> >> >
>> >> >> > APL25 USB Device Controller Incorrectly Interprets U3 Wakeup For
>> >> >> > Warm Reset
>> >> >> >
>> >> >> > APL27 USB 2.0 Timing Responsiveness Degradation
>> >> >> >
>> >> >> > APL50 System May Unexpectedly Shut Down When Software Requests a
>> >> >> > Reset
>> >> >> >
>> >> >> > APL53 Warm Reset May Result in System Hang or Unexpected System
>> >> >> > Behaviour
>> >> >> >
>> >> >> > Keep your eyes open and.. expect some more technical Atom CPU
>> >> >> > finding
Reply | Threaded
Open this post in threaded view
|

Re: SuperMicro A2SDi-4C-HLN4F

lists-2
Sun, 19 Aug 2018 08:34:43 +0000 Rupert Gallagher <[hidden email]>
> Troll here...

Rupert,

> Step out of your nerdiness, and step into  the real world with this scenario.

A troll mini speaker calls real world their empty words on a mailing list.
You are simply ignored.  In fact, you send total nonsense all around.  And
the basis of your miniature demanding pretences has no meaning whatsoever.

> A well-known vendor of server boards has an entite product line based on intel atom C3000 cpus.

Windows corporate & assembly shops targeting vendor.  "Producing", but not
actually carrying (relies on third party selling service and has a smaller
global SKU list) boards.  They frequently fail before, or near the limited
warranty.  At this point you better use desktop boards from other vendors.
Especially other CPU vendors outside Intel's dominating position of FAILS.

> Each product has a list of certified operating systems, including windows, freebsd, red hat linux enterprise, esxi.

You speak of the self proclaimed Enterprise segment, that is mostly vendor
locked hardware fault addressing extra cost.  You can not push your demand
and your wishes to have support where vendors try to exclude free systems.
Why don't you just ask direct the vendor instead, see what they give you?!

> The list does not include openbsd.

How are you helping this process, Rupert?  At this point, you are throwing
offence past simply comparing with sponsored software "other OS" versions.

> You spend a few days testing the server with various oss.

Minimum effort it is.  That is your own fault, everything else is ignored.

> It turns out that the server *works* with everything linux, including debian, proxmox, centos.

As in "works past the hardware faults", boots up.  You should rather use a
board that is more popular, has been tested out thoroughly, has seen wider
market adoption, bugs caught, with new revision of CPU and firmware fixes.

This vendor addresses hardware & firmware faults like the other enterprise
vendors, they DON'T past year two.  BIOS and BMC firmwares are not updated
after this even with the long term lifetime products, you are on your own!

> It also turns out that openbsd fails to recognise the devices, including the disks, and the keyboard.

You should really put some more effort into this to see how it progresses.

> Then you post its dmesg, because obsd folks really like it, together with a positive dmesg from centos.

So you dictate now which unsupported boards are in use?  You are a nobody.
Positive-negative is your opinion only, it doesn't matter past your mouth.

>  Someone intervenes spreading FUD about the server and flaming you are a troll.

You get what you ask for.  Says much about your original intent, to spread
negative abusive words.  You fail to disrupt anything, no one has the time
to read your boot up complaints with offensive language.  Get out of here.

Kind regards,
Anton Lazarov

> On Sun, Aug 19, 2018 at 00:06, <[hidden email]> wrote:
>
> > Sat, 18 Aug 2018 12:28:17 +0000 Rupert Gallagher <[hidden email]>  
> >> I am not complaining about obsd.
> >>
> >> I am complaining about your FUD on C3000 and your request for a
> >> post-installation dmesg. No installation is possible for lack of disk
> >> drivers. I do not develop/port drivers. The dmesgs I already posted
> >> are all I can share.
> >>
> >> The only way to run obsd at this time is via qemu/proxmox or
> >> vmware/esxi.  
> >
> > Rupert,
> >
> > You would not have to complain if you hadn't spammed your puny pretences.
> > The boards and processors have issues to be known, accompanying warnings.
> >
> > The dmesg you posted does not help anything. It looks like a poorly done
> > commercial for other operating systems that "supposedly" work. Only, you
> > have not been running anything past boot up. Accept it, you are a troll.
> >
> > Looks like you point out there are "drivers to be written", for your use.
> > I call this helpless whining. I told you how you can help this process..
> >
> > Get the boards in the hands of developers with OpenBSD who write drivers.
> > And first kindly ask about it whether it is possible to fit in busy work.
> >
> > Kind regards,
> > Anton Lazarov
> >  
> >> On Sat, Aug 18, 2018 at 12:55, <[hidden email]> wrote:
> >>  
> >> > Sat, 18 Aug 2018 08:29:33 +0000 Rupert Gallagher <[hidden email]>  
> >> >> Orher architectures are explicitly known for being bugged, but this
> >> >> did not stop the developers.  
> >> >
> >> > Hi Rupert,
> >> >
> >> > Enough already, just post a full dmesg when you can. The hardware is not
> >> > perfect, watch out for unfixed CPU bugs as these are soldered down chips.
> >> >
> >> > They also include a lot more than the CPU where most of the problems are.
> >> > So keep an eye on these boards and be a bit more impartial and objective.
> >> >
> >> > It would be so nice if you could ship new boards like these to developers
> >> > or donate some of your excess hardware costs so that they get the boards.
> >> >
> >> > That might be much more helpful than complaints over the installer dmesg.
> >> >
> >> > Kind regards,
> >> > Anton Lazarov
> >> >  
> >> >> On Fri, Aug 17, 2018 at 22:09, <[hidden email]> wrote:
> >> >>  
> >> >> > Fri, 17 Aug 2018 16:48:52 +0000 Rupert Gallagher <[hidden email]>  
> >> >> >> This is a really nice board at a really nice price, and you should
> >> >> >> stop scaring people off. These Atoms are cleaner than both Cores and
> >> >> >> Xeons, and AMDs have have their fair share of problems. Citin g
> >> >> >> C2000 against the latest C3000 is FUD. You should come clean first,
> >> >> >> and cite a comparable server board at a comparable cost and clear of
> >> >> >> bugs.  
> >> >> >
> >> >> > Rupert,
> >> >> >
> >> >> > The publication and processor specifications are about the C3000 CPU bug.
> >> >> > Similar hardware flaws affect the new C3000 CPUs, same microarchitecture.
> >> >> > These Atoms are also fully affected by Intel HW flaws Spectre & Meltdown.
> >> >> >
> >> >> > With NONE real hardware understanding, you have to rely upon vendor spec.
> >> >> > With NONE single dmesg after installation, you can NOT advise this board.
> >> >> > With NONE real OpenBSD experience, you better ask the developers instead.
> >> >> >
> >> >> > You are speculatively advertising yet unsupported boards as feature sets.
> >> >> > Not interested reading anything that pops in your visibility on the list.
> >> >> > It takes very little effort to have some opinion and I do NOT want yours.
> >> >> >
> >> >> > Kind regards,
> >> >> > Anton Lazarov
> >> >> >  
> >> >> >> On Fri, Aug 17, 2018 at 17:38, <[hidden email]> wrote:
> >> >> >>  
> >> >> >> > Fri, 17 Aug 2018 11:15:32 +0000 Rupert Gallagher
> >> >> >> > <[hidden email]>  
> >> >> >> >> FUD
> >> >> >> >>
> >> >> >> >> Sent from total and utter ignorance.  
> >> >> >> >
> >> >> >> > Rupert,
> >> >> >> >
> >> >> >> > Plus the dmesg you posted is by the RAM disk installation OpenBSD
> >> >> >> > kernel. You Should complete an actual OpenBSD installation to get
> >> >> >> > the full dmesg.  
> >> >> >> >> On Fri, Aug 17, 2018 at 12:24, <[hidden email]> wrote:
> >> >> >> >>  
> >> >> >> >> > Fri, 17 Aug 2018 06:23:23 +0000 Rupert Gallagher
> >> >> >> >> > <[hidden email]>  
> >> >> >> >> >> Using the serial over lan console, the keyboard keeps working,
> >> >> >> >> >> and it is possible to enter the shell. Using the standard
> >> >> >> >> >> console, the keyboard does not work. There is plenty of drivers
> >> >> >> >> >> that need to be written. No, I do not write/port drivers.
> >> >> >> >> >>
> >> >> >> >> >> By comparison, both freebsd and centos linux just work out of
> >> >> >> >> >> the box. Both dmesgs are attached as reference.
> >> >> >> >> >>
> >> >> >> >> >> The board is really nice.
> >> >> >> >> >>
> >> >> >> >> >> On Thu, Aug 16, 2018 at 18:04, Rupert Gallagher
> >> >> >> >> >> <[hidden email]> wrote:  
> >> >> >> >> >> > https://www.supermicro.com/products/motherboard/atom/A2SDi-4C-HLN4F.cfm 
> >> >> >> >> >
> >> >> >> >> > Hi misc@,
> >> >> >> >> >
> >> >> >> >> > This one is much quieter, the vendors have reacted to the
> >> >> >> >> > previous flaw. As nice as it can get.. before trusting
> >> >> >> >> > non-technical reviews, be aware:
> >> >> >> >> >
> >> >> >> >> > https://en.wikipedia.org/wiki/Goldmont#Erratum
> >> >> >> >> > https://www.servethehome.com/another-atom-bomb-intel-e3800-bay-trail-atom-vli89-bug/
> >> >> >> >> > https://www.google.com/search?q=intel+c3000+bug
> >> >> >> >> >
> >> >> >> >> > Recurring issue - similar problems were discussed noisily
> >> >> >> >> > industry wide:
> >> >> >> >> >
> >> >> >> >> > https://en.wikipedia.org/wiki/Silvermont#Erratum
> >> >> >> >> > https://www.servethehome.com/intel-atom-c2000-series-bug-quiet/
> >> >> >> >> > https://www.servethehome.com/intel-atom-c2000-c0-stepping-fixing-the-avr54-bug/
> >> >> >> >> > https://www.google.com/search?q=intel+c2000+bug
> >> >> >> >> >
> >> >> >> >> > As with the previous generation processors and main boards,
> >> >> >> >> > Intel FAILS. Wait for updated CPU revision boards, consult techs
> >> >> >> >> > and prepare for RMA.
> >> >> >> >> >
> >> >> >> >> > Kind regards,
> >> >> >> >> > Anton Lazarov  
> >> >> >> >
> >> >> >> > "Similar to previous Silvermont generation design flaws were found
> >> >> >> > in processor circuitry resulting in cease of operation when
> >> >> >> > processors are actively used for several years. Errata named
> >> >> >> >
> >> >> >> > APL47 System May Experience Inability to Boot or May Cease Operation
> >> >> >> >
> >> >> >> > and
> >> >> >> >
> >> >> >> > APL48 System May Experience Non-functioning GPIO Weak Pull-up
> >> >> >> > Circuitry
> >> >> >> >
> >> >> >> > were added to documentation in June 2017 stating:
> >> >> >> >
> >> >> >> > The Low Pin Count (LPC), Real-Time Clock (RTC), Secure Digital (SD)
> >> >> >> > card and General-Purpose Input/Output (GPIO) interfaces may stop
> >> >> >> > functioning.
> >> >> >> >
> >> >> >> > The documents cited in the Wikipedia article, see for current
> >> >> >> > versions.. that apply to the later CPU stepping models, if any, as
> >> >> >> > time progresses:
> >> >> >> >
> >> >> >> > https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/pentium-celeron-n-series-j-series-datasheet-spec-update.pdf
> >> >> >> > https://www.intel.com/content/dam/www/public/us/en/documents/specification-updates/atom-c3000-family-spec-update.pdf
> >> >> >> >
> >> >> >> > APL47 Problem: Under certain conditions LPC, SD card and RTC
> >> >> >> > circuitry may stop functioning in the outer years of use.
> >> >> >> >
> >> >> >> > APL48 Problem: When platform drives the GPIO pin low, GPIOs
> >> >> >> > programmed with weak pull-up circuitry may not maintain a value
> >> >> >> > above VIH when not actively driven in outer years of service."
> >> >> >> >
> >> >> >> > Some other "interesting" errata:
> >> >> >> >
> >> >> >> > APL23 HD Audio Recording May Experience a Glitch While Opening or
> >> >> >> > Closing Audio Streams
> >> >> >> >
> >> >> >> > APL25 USB Device Controller Incorrectly Interprets U3 Wakeup For
> >> >> >> > Warm Reset
> >> >> >> >
> >> >> >> > APL27 USB 2.0 Timing Responsiveness Degradation
> >> >> >> >
> >> >> >> > APL50 System May Unexpectedly Shut Down When Software Requests a
> >> >> >> > Reset
> >> >> >> >
> >> >> >> > APL53 Warm Reset May Result in System Hang or Unexpected System
> >> >> >> > Behaviour
> >> >> >> >
> >> >> >> > Keep your eyes open and.. expect some more technical Atom CPU
> >> >> >> > findin  

Reply | Threaded
Open this post in threaded view
|

Re: SuperMicro A2SDi-4C-HLN4F

phessler
On 2018 Aug 20 (Mon) at 05:29:50 +0300 (+0300), [hidden email] wrote:
:You get what you ask for.  Says much about your original intent, to spread
:negative abusive words.  You fail to disrupt anything, no one has the time
:to read your boot up complaints with offensive language.  Get out of here.

wrant, Shut up.  You are routinely abusive and not helpful.  Don't
lecture people on tone, or how useful someone is.  You have always
failed in that comparision.

Don't respond, just shut up.

Reply | Threaded
Open this post in threaded view
|

Re: SuperMicro A2SDi-4C-HLN4F

Rupert Gallagher
In reply to this post by Rupert Gallagher
<[hidden email]> wrote:

> This vendor addresses hardware & firmware faults like the other enterprise vendors, they DON'T past year two. BIOS and BMC firmwares are not updated after this even with the long term lifetime products, you are on your own!

On bios and ipmi updates, you can download and apply them yourself. For advanced bios updates, there is a licence you can purchase, as you do with Dell and HP.

On warranty, they are in line with the industry standards:

https://www.supermicro.com/support/Warranty/

On open source, they work with the community, well enough to have a cetified list of compatible systems:

https://www.supermicro.com/support/faqs/os.cfm

When you are in business, you do not want to go back to the drawing board each time. You need a platform where to build your own services. If you have to develop disk and keyboard drivers, and you are told off from the support mailing list, then the OS is worth nothing to you. OpenBSD is not ready for enterprise.

Your faithfull troll.
Reply | Threaded
Open this post in threaded view
|

Re: SuperMicro A2SDi-4C-HLN4F

Mihai Popescu-3
In reply to this post by Rupert Gallagher
> OpenBSD is not ready for enterprise.

Fantastic intuition here [1].

[1] https://marc.info/?l=openbsd-misc&m=153385998722067&w=2

Reply | Threaded
Open this post in threaded view
|

Re: SuperMicro A2SDi-4C-HLN4F

Rupert Gallagher
In reply to this post by Rupert Gallagher
Popescu, obsd does not always work. Hiding the bugs, instead of showing them on git issues or bugzilla, it does not make obsd any better.

Sent from ProtonMail Mobile
Reply | Threaded
Open this post in threaded view
|

Re: SuperMicro A2SDi-4C-HLN4F

Stuart Henderson
In reply to this post by Mihai Popescu-3
On 2018-08-26, Mihai Popescu <[hidden email]> wrote:
>> OpenBSD is not ready for enterprise.
>
> Fantastic intuition here [1].
>
> [1] https://marc.info/?l=openbsd-misc&m=153385998722067&w=2
>
>

Would you mind fixing your thread headers please? My killfile entry for
threads with 'protonmail.com' in references doesn't work very well otherwise ;)

Reply | Threaded
Open this post in threaded view
|

Re: SuperMicro A2SDi-4C-HLN4F

Alexander Hall
In reply to this post by Rupert Gallagher


On August 25, 2018 1:59:55 PM GMT+02:00, Rupert Gallagher <[hidden email]> wrote:

><[hidden email]> wrote:
>
>> This vendor addresses hardware & firmware faults like the other
>enterprise vendors, they DON'T past year two. BIOS and BMC firmwares
>are not updated after this even with the long term lifetime products,
>you are on your own!
>
>On bios and ipmi updates, you can download and apply them yourself. For
>advanced bios updates, there is a licence you can purchase, as you do
>with Dell and HP.
>
>On warranty, they are in line with the industry standards:
>
>https://www.supermicro.com/support/Warranty/
>
>On open source, they work with the community, well enough to have a
>cetified list of compatible systems:
>
>https://www.supermicro.com/support/faqs/os.cfm
>
>When you are in business, you do not want to go back to the drawing
>board each time. You need a platform where to build your own services.
>If you have to develop disk and keyboard drivers, and you are told off
>from the support mailing list, then the OS is worth nothing to you.

- That there is no current support, does not mean there will never be.
- misc@ is not a support mailing list. In fact, there is none. That does not, however, mean that you can't get help from there.

>OpenBSD is not ready for enterprise.

By some definition of "for enterprise", sure.

/Alexander

>
>Your faithfull troll.